Last week we stopped ordering the comparison on our B2B tech PR page by our own index. We put our profile first and listed the other 12 agencies alphabetically. In week 8 of our UK B2B Tech PR AI Visibility Index, ChatGPT changed which agencies it picked from the page. It named us no more often than before.
In week 7 ChatGPT cited our comparison page in 40 answers, more than any other source, and named us in 28. The table on the page followed the order of our index, where we were 9th. ChatGPT appeared to build its shortlist from the top of that table and then check each agency on its own website. Our page was handing ChatGPT a ranking with eight agencies above us.
Everything else stayed the same, including the URL and what we say about each agency.
ChatGPT cited the page in 52 answers this week, up from 40, and its picks changed. Hotwire had been at the top of our table. Last week ChatGPT named it in 32 of the 40 answers that cited the page. This week it named Hotwire in 25 of 52. Babel PR went from 23 of those answers to 37, and CommsCo from 13 to 26.
The effect carried through to ChatGPT's own ranking. Across all its answers it now puts Babel PR first, in 49 answers, and Hotwire sixth, in 39. Last week Hotwire was first, in 51. ChatGPT's answers to our check questions held, so we think the change came from the page.
The new order doesn't explain everything. Archetype comes straight after us on the page, and ChatGPT named it in 28 answers both weeks. What we can say is that the agencies at the top of the old table lost the lead that table gave them.
ChatGPT named us in 27 answers, against 28 last week, and moved us from 10th to 13th in its own ranking. It seems to treat the author of a comparison as the source of the list and leaves the author off the shortlist it builds from it. Our name appeared only as a label on a source in 43 answers across the six assistants, which we don't count as a mention.
Grok named us in 56 answers, the same as last week. Perplexity named us in 33, against 35, and Gemini in 17, against 16. Google AI Mode named us in 15, up from 11, and AI Overviews in 14, up from 11. The check answers changed on Gemini, Grok and both Google surfaces, so part of their movement comes from the assistants themselves.
We stayed 9th overall, on 3.9% of the weighted share of mentions. Babel PR leads the index for the first time, on 7.1%, and seven tenths of a point separate the top six. Our comparison page was the most-cited page in the study again, in 175 answers.
Once we leave out the answers that cite our own site, we're 27th, down from 18th, and named in 12 answers. Other agencies' lists are gaining ground at the same time. Jamco PR's list of UK tech PR agencies was cited in 39 answers after none last week, and Pearl Lemon PR's in 31, up from 7.
We've now learned two things about comparison pages. The order of a list changes which agencies an assistant picks from it, and the author of the list gets little of the benefit. So our next step is to be named on lists and review sites that other people publish.
The full ranking and every week so far are on the index page. To see which pages the assistants read and recommend from in your own category, ask us for a free AI visibility snapshot.